Dosimetric evaluation of skin collimation with tungsten rubber for electron radiotherapy: A Monte Carlo study

Abstract Purpose Skin collimation provides a sharp penumbra for electron beams, while the effect of bremsstrahlung from shielding materials is a concern. This phantom study was conducted to evaluate the safety and efficacy of a real‐time variable shape rubber containing‐tungsten (STR) that can be placed on a patient’s skin. Methods Electron beam profiles were acquired with the STR placed on a water‐equivalent phantom and low melting‐point alloy (LMA) placed at the applicator according to commonly used procedures (field sizes: 20‐ and 40‐mm diameters). Depth and lateral dose profiles for 6‐ and 12‐MeV electron beams were obtained by Monte Carlo (MC) simulations and were benchmarked against film measurements. The width of the off‐axis distance between 80% and 20% doses (P 80‐20) and the maximum dose were obtained from the lateral dose profiles. Bremsstrahlung emission was analyzed by MC simulations at the depth of maximum dose (R 100). Results The depth dose profiles calculated by the MC simulations were consistently within 2% of the measurements. The P 80‐20 at R 100 for 20‐ and 40‐mm diameters were 4.0 mm vs. 7.6 mm (STR vs. LMA) and 4.5 mm vs. 9.2 mm, respectively, for the 6‐MeV electron beam with 7.0‐mm‐thick STR, and 2.7 mm vs. 5.6 mm and 4.5 mm vs. 7.1 mm, respectively, for the 12‐MeV electron beam with 12.0‐mm‐thick STR. A hotspot was not observed on the lateral dose profiles obtained with the STR at R 100. The bremsstrahlung emission under the region shielded by the STR was comparable to that obtained with the LMA, even though the STR was placed on the surface of the phantom. Conclusions Skin collimator with STR provided superior dosimetric characteristics and comparable bremsstrahlung emission to LMA collimator at the applicator. STR could be a new tool for the safe and efficient delivery of electron radiotherapy.


| INTRODUCTION
In electron radiotherapy, lead and low melting-point alloy (LMA) are widely used to shape the irradiation field because of their excellent shielding ability against electron beams. 1-3 However, the shortcoming of shielding materials made of lead or LMA equivalent includes toxicity to the human body. 4,5 In Japan, lead and LMA are commonly placed at the applicator to shape the irradiation field of the electron beam. 6 However, an irradiation field shaped at the applicator has the disadvantage of a large penumbra compared with that shaped by a skin collimator. 7 In a previous study, we reported that shielding materials containing tungsten, such as tungsten functional paper (TFP) 7-12 and tungsten-containing rubber (TCR), [13][14][15] can provide sufficient radiation shielding against therapeutic electron beams and diagnostic xrays and γ-rays. In electron therapy, TCR offers the advantages of being nontoxic to the human body and being easy to shape the irradiation field. Recently, we developed real-time variable shape rubber containing-tungsten (STR) (Hayakawa Rubber Co., Ltd. Hiroshima, Japan) and demonstrated that the transmission rates reached plateau values at STR with thicknesses of ≥7.0 and ≥12.0 mm for 6-and 12-MeV electron beams, respectively. 16 A major advantage of the STR is that its shape can be changed by hand at a temperature of approximately 60°C, with it then maintaining this shape at room and body temperatures. 16 A dynamic mechanical analysis showed that the ratio of viscosity to elasticity was 1.016 at the temperature of approximately 60°C, which is higher than the value of 0.6 required for shaping by hand. 16 Therefore, the STR has the potential to facilitate flexible and immediate shaping of the irradiation field of electron beams.
Skin collimation is expected to produce a sharper penumbra 17 and provide higher quality dose distribution in small-field electron beam therapy. 17,18 On the other hand, the effect of bremsstrahlung from the shielding material is a concern by skin collimation. 11 Kamomae et al. reported a slight increase in photon components at the exit plane of TFP as a chest wall protective disc in intraoperative electron radiotherapy for breast cancer treatment. 12 However, it remains unclear whether the effect of bremsstrahlung from skin collimation is greater than that of shielding by conventional applicators in treatment for superficial lesions. Furthermore, for small-field electron beams, the depth of the maximum dose and the sharpness of the dose fall-off are different from those of a broad beam distribution because of the loss of lateral scatter equilibrium (LSE). 19 Therefore, the dosimetric characteristics of small-field electron beams for skin collimation with STR must be clarified.
The purpose of this study was to evaluate the occurrence of bremsstrahlung to demonstrate the safe delivery of electron radiotherapy using an STR skin collimator, making comparisons with a conventional LMA collimator placed at the applicator. Monte Carlo (MC) simulations were used for these evaluations. In addition, lateral dose profiles and percent depth doses (PDDs) were also examined to investigate the dose distribution of small-field electron beams shaped by STR.

2.A | Experimental setup
Measurements of PDDs and lateral dose profiles in a water-equivalent solid water phantom (GAMMEX, Wisconsin, USA) were performed with Gafchromic EBT3 films (ISP, Wayne, NJ, USA) for smallfield electron beams shaped with LMA or STR (Fig. 1). To shape the irradiation field, the LMA was placed at the applicator, whereas the STR was set on the surface of the solid water phantom. The density of the LMA was 9.8 g/cm 3 and the element ratio (wt%) was In: 2.0%, Sn: 20.0%, Pb: 32.0%, and Bi: 46.0%. The density of the STR was 7.3 g/cm 3 and the element ratio (wt%) was C: 5.5%, H: 0.9%, O: 1.4%, and W: 92.2%. The characteristics of the STR were described in detail in a previous report 16  A dose calibration curve for the EBT3 film was prepared by irradiating the film with doses from 0.0 to 3.0 Gy, with an irradiation field size of 10 cm × 10 cm at the depth of the maximum dose in a solid water phantom. The exposed films were scanned using an ES-G10000 (Epson Corp., Nagano, Japan) with a resolution of 72 dotsper-inch in 48-bit RGB color images. All film measurements were analyzed from the red channel using ImageJ version 1.52a (National Institutes of Health, Bethesda, MD).

2.B | Measurement of the dose profiles
The film was set up parallel to the beam central axis (CAX) for PDDs and lateral dose profiles measurements. The lateral dose profiles at the surface were measured by arranging the film so that it was perpendicular to the beam axis at the phantom surface. The PDDs were normalized to the maximum dose. The depths of the maximum dose (R 100 ), 90% dose (R 90 ), 80% dose (R 80 ), and photon contamination dose (D x ; beyond the maximum range of electrons 19 ) were obtained from the PDD curves. The lateral dose profiles at each depth were normalized by the dose at the CAX. The PDDs and lateral dose profiles of LMA and STR were compared for each field size and electron energy. The maximum doses within the irradiated field and the penumbra (as the width of the off-axis distance from 80% to 20% dose levels, P 80-20 ) were evaluated at the surface, R 100 , R 90 , and R 80 . 16

2.C | Monte Carlo simulations
The particle and heavy ion transport code system (PHITS) (version 2.74, Japan Atomic Energy Agency, Tokai, Japan) was used to simulate the behavior of the electron beam. PHITS can deal with the transport of nearly all particles (i.e., neutrons, protons, heavy ions, photons, and electrons) over wide energy ranges in three-dimensional modeling systems. 20 The Electron Gamma Shower version 5 (EGS5) algorithm was used for electron and photon transport in this study. The calculation parameters included a cut-off for electron and positron kinetic energy of 0.1 MeV, photon cut-off energy of 0.01 MeV, and 1.0 × 10 8 source electrons. The number of particle histories was determined so that the statistical uncertainties evaluated as the standard error would be within ±1%, ±2%, and ±2% for each PDD, lateral dose profile, and spectral distribution, respectively. 9 In this study, a simulation was performed using phase space files because the in-head information of the TrueBeam linac is not disclosed. The phase space files for the TrueBeam linac were provided by the vendor in an International Atomic Energy Agency (IAEA) compatible format, and the details of these files are described by Rodrigues et al. 21 The geometries in the MC simulations reproduced the experimental setups shown in Fig. 1. 100.1% and 100.2%, respectively ( Fig. 4b and 4f).

3.B | Monte Carlo simulation modeling
The PDDs (Fig. 2) and lateral dose profiles with MC simulation

| DISCUSSION
In this study, we evaluated the dosimetric characteristics of smallfield electron beams for skin collimation with STR, and the electron radiotherapy with the STR placed on the surface of a patient could be performed safely and effectively.
The PDDs (Fig. 2 and Table 1) were almost equal between LMA and STR, which indicates that the dosimetric characteristics of the depth dose profile were not affected by the shielding material and position. This result was in agreement with our previous report 16 demonstrating that the method was effective even for a small-field. The STR also provided a sharp penumbra by shielding close to the patient, which has advantages when protecting adjacent critical structures from small-field electron therapy. The maximum doses around the edge of the irradiation field with STR at the phantom surface were higher than those with LMA (Figs. 3a, 3e, 4a, and 4e) because of the effects of electron scattering from the edge of the Off-axis distance (mm) STR. 17 However, for a 40-mm diameter, the relative magnitudes of those high doses were approximately 80% and 90% of the prescribed doses with 6-and 12-MeV electron beams, because the PDD values at the surface were approximately 75% and 85%, as shown in Fig. 2. 16 In addition, even though thick STR is needed (e.g., 12 mm), it is easy to cut pieces out of the material by hand, and it provides a sharp dose falloff at the edge of the irradiation field.
Although a variable thickness, especially at the edges, may cause a high dose due to penetration and scatter, 11,16 in this case, the high dose was at most 90% of the prescribed dose. Therefore, we T A B L E 1 Dosimetric characteristics of PDDs for 20-and 40-mm diameters with LMA and 7.0-and 12.0-mm-thick STRs for 6-and 12-MeV electron beams.  conclude that manual shaping of STR was not an issue in relation to beam divergence. The high dose region decreased as the depth increased because the scattered electrons were low energy. 17 The spectral distribution analysis showed that the generation of bremsstrahlung from the shielding material 12 was not a clinical concern. The bremsstrahlung from the STR was lower or comparable to that from the LMA, except for the low energy region (<3 MeV) for the 12-MeV electron beam under the shielded area. However, with the STR, the increase in photon components within this energy region was at most 2% ( Fig. 5e-5h), while the standard error of the MC simulation was ±2%, indicating the difference was not significant. Furthermore, photon contamination was comparable between STR and LMA, as shown in Table 1. On the other hand, the electron distribution of STR correlated with LMA in the CAX, and was comparable to or less than that of LMA under the shielded area; hence, STR could provide the safe delivery in clinical practice.
The results of our experiments were consistent with those of Perez et al. 22 who reported results for lead-based shielding. As the placement of the shielding material directly on the skin reduces the spread of the penumbra, we expect that STR will reduce the additional margin and exposure to normal tissues in comparison with previous methods involving the placement of LMA at the applicator.
The STR also has the advantage that it can be shaped in real-time, thereby facilitating quick and individualized electron radiotherapy for each patient, such as for the treatment of keloid scars, where it is important to start the treatment within 7 h after excision to decrease recurrence rates. 23 A real-time shaping procedure could proceed as follows: (a) the STR wrapped in clingfilm is heated in a 600 W microwave oven for 1 min, (b) it is rolled out to a sheet of uniform thickness, (c) the irradiation field is cut out of the STR using a shaped cutter or scissors. In addition, STR can adapt to the patient's respiratory motion and the rounded contours of the patient's body; it can be placed directly on the body surface because it is nontoxic. 16 STR could be useful for lesions of the ear, nose, breast, belly, and those around the eye, and it could also be applied in intraoperative electron radiotherapy, 12 in electron grid therapy, 9 and as an eye shield 24 instead of conventional devices.

| CONCLUSION
The STR skin collimator provided superior dosimetric characteristics and comparable the occurrence of bremsstrahlung to the conventional LMA collimator at the applicators. STR, which can be placed on the surface of the patient, is a promising new tool to aid the safe and effective delivery of small-field electron radiotherapy.

ACKNOWLEDGMENTS
This work was supported in part by the Japan Society for the Promotion of Science (JSPS) KAKENHI grant number 19K08211. We thank Karl Embleton, PhD, from Edanz Group (https://en-author-se rvices.edanzgroup.com/) for editing a draft of this manuscript.